A Workflow for Real Experts
AI has become one of those topics where everyone has an opinion, but very few people have a workflow.
That became clear to me recently when my wife was asked to help shape an AI strategy for the bank she works for. What stood out wasn’t that she had a formal background in artificial intelligence. She didn’t. What stood out was where her confidence came from.
Most of what she brought into that conversation was learned by actually using AI day to day — inside her health and wellness work, inside our YouTube workflow, and inside the messy reality of building something that has to function, not just sound impressive.
Watching that process forced me to slow down and reflect on how we use AI, too. Where it’s genuinely helped us think better. Where it’s saved time. And just as importantly, where it’s actively made things worse.
Because the truth is, we didn’t arrive at a clean system by reading think pieces or copying prompts. We arrived there through trial, frustration, and a few uncomfortable realizations about what AI is good at — and what it absolutely shouldn’t be trusted to do.That reflection led to a simple conclusion: AI doesn’t fail because it’s replacing human thinking. It fails when it’s asked to lead before the human has decided anything.
The Real Issue Isn’t AI — It’s Outsourced Judgment
When AI is asked to decide what a piece of content should be about, what angle to take, or what actually matters, the result is usually forgettable. Not because the tool is broken, but because judgment isn’t a computational task.
Judgment comes from experience. From taste. From context. From having watched ideas succeed and fail in the real world.
That’s the part experts are paid for. And it’s the part AI should never be responsible for.
When that responsibility is handed over too early, the output may sound polished, but it lacks shape. There’s no conviction behind it because no one ever made a decision worth standing behind. You can feel that absence immediately as a reader.
The problem isn’t that AI is too powerful.
It’s that it’s being used out of order.
What Using AI in the Real World Taught Us
Across both business work and content work, the pattern was consistent.
When we tried to use AI to figure out what to say, the output felt flat. Technically correct, but hollow. Clean, but generic. Fast, but disconnected from anything we actually believed.
When we used AI after we had already decided what mattered, everything changed.
It helped us pressure-test ideas we already believed in. It clarified language without diluting intent. It surfaced blind spots we hadn’t considered. It gave us alternative framings without pulling us away from our point of view.
The difference wasn’t the tool.
It was the order.
That’s the part that gets missed in most AI conversations, especially in business settings. People jump straight to efficiency and capability without addressing the more important question: Where does human judgment end, and where should AI begin?
A More Useful Way to Think About AI
The most effective way I’ve seen AI used isn’t as a writer or a creator. It’s as a thinking partner that shows up after intent is clear.
In practice, that means the human does the hard work first.
You decide what the payoff is.
You clarify what tension actually matters.
You commit to a point of view.
You choose what not to say.
Only then does AI enter the workflow.
At that stage, it becomes genuinely useful. Not because it adds originality, but because it helps test clarity. It can challenge weak assumptions, expose gaps in logic, or highlight places where language is getting in the way of meaning.
AI works best when it’s responding to thinking, not replacing it.
Where AI Fits — and Where It Doesn’t
In a healthy workflow, AI supports execution, not direction.
It’s useful for pressure-testing an idea you already believe in. It’s helpful when you want to explore alternate ways of explaining the same concept. It’s valuable when clarity matters more than speed.
It becomes a problem when it’s asked to generate insight from scratch, decide what the message should be, or manufacture conviction that doesn’t exist yet.
That’s when content starts to sound interchangeable. Not because AI is flattening creativity, but because no one ever stepped in to shape it.
Why Authority Still Belongs to Humans
There’s a growing fear that AI will flatten expertise and make everyone sound the same. In practice, the opposite is happening.
As AI-generated content floods the internet, human judgment becomes more visible, not less. Clear thinking stands out faster. So does lived experience. So does someone who knows what matters and is willing to say it plainly.
AI can accelerate output.
It cannot manufacture authority.
Authority comes from making decisions, standing behind them, and repeating them consistently enough that people recognize your point of view. AI can support that process, but it cannot lead it.
The Rule I Use Now
I follow one simple rule.
If AI is deciding what I’m saying, I stop.
If AI is helping me say it more clearly, I continue.
That distinction has kept our work coherent, defensible, and human — even as the tools keep changing around us.
Want to Go Deeper?
If this way of thinking resonates, there are a few ways to continue the conversation.
I share ongoing reflections like this as I refine our workflows and thinking in real time. You can follow those discussions on Reddit.
And if you want help building an AI-enabled workflow that sharpens your thinking instead of diluting it, you can book a strategy call with me. That work is always tailored, never templated.No hacks. No autopilot.
Just clearer thinking, applied consistently.